Comparing the stability of successional transient models in kelp forests and barrens (south‐east Pacific): Implications for conservation monitoring